I am working on a search engine. Users can search for other users givin specific parameters (such for example age, gender, city, hair color). He is also able to sort the results in different ways.
On the search result page you are able to see 10 profiles and if you click on them, you get to another page (profile page) where you have more information on that single user.
The user who's now on a profil page looking at a specific user should now have to possibility to go back and forth to the next or the previous user from the search result. So you would have to store the specific parameters the user chose and you would also have to run the whole search query over and over again as the user continues to go from one user to the next. This doesn't look nice to me and I thought about storing all user ids temporarily in a session (max 200 to previous users and max 200 to upcoming users) to avoid store the parameters and run the search query every time the user vists another profile page.
Is this a good way or do you have other solutions?
Best regards and thanks in advance for your help,
Freddy
You could simply store the id's as an array in a session, and then put next/previous links to the next/previous items in the array.
Wit small number of total matches (under 200 for example), storing IDs in the session is perfectly fine. If you potentially expect a large number of results in the search (in thousands), then a better solution may be to implement a temporary table for the searches. When a search is performed, the IDs and names of the matches are stored in a temporary table against the session ID of the user. You'd need to also have a script that runs periodically and cleans this table of old data.
Related
I find several partial solutions in answers on this question, but common answer seems to be absent.
So, I have a table users with columns user_id, user_name. On each static page of website I want to display all user names of users who currently view this page.
Should I have a table views with columns user_id, webpage_link?
If yes, when I shall update data in column webpage_link? How to connect code from following answer with mysql database Is there a way to detect if a browser window is not currently active? ? (If it is ok for this purpose.)
To make updates very often is not very good. So, the user can view several pages (for example, in 2 or more tabs). What type of webpage_link column shall be in this case?
With every http request, you get a $path variable. if you also have a logged in user, you can store which page this user requested last (e.g. in a table like you described, but only storing the relative path).
You update this information on a per-request-basis in some sort of front-controller. (just make sure you put it where it is called for every authenticated page). When the users session times out, you remove the row of that user from the table.
this case is a little more difficult. you could store the last n pages/paths the user has requested and leave the rest as above. You don't have to change the table structure for that, just allow for multiple rows per user. (the combination user_id+path should be unique, though)
Hope that helps to get you started
I'm currently looking for the BEST and SMOOTHEST way to get the active users on my website using a particular page. (purpose: think game lobby).
I was thinking about storing each visitor in a table, with their IP-adress as primary key in the database, together with a lastAction field - which stores a timestamp of their last action.
Gathering the active users would then simply just be to gather all rows where the last action was within the correct amount of time.
The problem: people on the same network (same IP) would be identified as the same user - it'd be hard to tell if they actually were a real person or just a page refresh with an old session. You get the idea.
I'm looking for smoother ways to separate users, or preferably, get away from the whole IP-adress part. Any ideas or suggestions would be appreciated.
Why not use a session based algorithm to determine the count of 'active users'? If you setup your session handler to write to a database (instead of flatfile), you can run queries against rows within the table to determine how many visitors and which page they are currently viewing.
You can read about how to switch your session handler here, but it does appear to be outdated. I suspect more googling will result in newer implementations.
I am programming a Blog. I am interested how to add a counter to the side to see how many people read the different articles.
I understood that if you have a Website counter, than you would normally save the IP and the Date in the database so you only count the person once.
But how would I make sure that a person reading a specific article is only counted once for that article?
The database would turn very big if I would save each Visitor IP for each Article. It came now into my mind to maybe create a Session for every Visitor. Within this Session I save each ArticleID the Person read. So before I update the counter in the database, I check if it already exists in Session - if not increment counter and save in Session otherwise dont.
Is this the right approach or is there another way to do it?
It depends on which data you want to save. You can save:
Number of hits (count all people/bots/... who opened the pages), this makes fake impression of readers, because doing a refresh will count me twice.
Number of unique visitors, create a session for the user. and increment the counter only once in the session. This can be improved by setting permanent cookie, but looks like overkill for me. and by the way, I can open the article and don't read it!
Number of people interact with the article. You can put some javascript capture (mouse scrolling, clicks on article text, ...) , split text into parts with "read more" ... etc.
If I were you, I will name it "number of views", and use solution 2. And combine this with google analytics (or similar solution) to know more details about visitors.
Google analytics is better way
But if you want to do it by programming then follow the steps.
i) In your table from the data of particular article comes just add one column no_of_view or whatever you want to give..
ii) In the starting of page write down the code of php for ::
from the url you can get the id of current article.
then from select query get the value of column no_of_view for particular ID (article)
then fire a query for update that record for no_of_view column with +1
If you want to get that more accurate then add ipaddress column and date also....
I prefer to use Google Analitycs API for such tasks. It' hard to implement for the first time - but once made it will help you on any project.
You can read here - http://code.google.com/intl/en/apis/analytics/docs/gdata/home.html
I would set a cookie on the user's browser and send it to the server. If the cookie is new, it's a new user reading the article for the first time.
I think, you can execute an ajax function, when user click on article to update column in database table field name 'number_of_click' as per article id and other relevant reference
and show it by simple query 'SELECT no_of_click FROM article WHERE id='{$individual_article_id} ....'';
Whats the best way to keep track of how many users and guests are online? Im making a forum for fun and learning
Right Now I have a 2 fields in the users table called is_online and last_access_time.
If current time is 5 minutes or more than last_access_time i set is_online it to zero. And if the signed in user refreshes browser i set it to 1.
But what about guests? I wanna keep track on how many guests are on also
Another thing that would be really cool is to show what page the user is viewing. and on the page, forum thread for example, 5 guests, Homer and Shomer are viewing this page. But how should i structure this? Hmm maybe i should make another question for that.
i dont know what i should do
What do you suggest?
I'd use cookies for this. Set a cookie when the user enters (checking first to make sure one doesnt exist). Easy way to generate a unique id for that user is to hash their IP plus the current time.
$id = md5($_SERVER['REMOTE_ADDR'] . time());
Store that id in your database and use that to reference
You can check what page they are viewing by grabbing either $_SERVER['PHP_SELF'] or $_SERVER['REQUEST_URI'] near the top of your php source. Store that in the table. I'd take a look at php.net's explanation of whats stored in the _SERVER global, as it should help out quite a bit if you find that you need more then just the document they are on (ex index.php). Found here.
You may need to pull apart of the query string that was used to access that page, parse out the variables to determine the page they are requesting. Either way, this could all be done through cookies, or just use a single cookie to store the unique id and use your table for storing everything else.
You cannot know for certain which page a user is viewing, but you can keep track of which page they last viewed. Every time you deliver a page to a user, record that page's path in a database row associated with them. Voila.
To keep the number of guests, I suggest tracking the number of distinct unauthenticated IP/HTTP-User-Agent combinations seen on a certain page in the last X minutes.
I found this article on Web Monkey that might help you.
http://www.webmonkey.com/2010/02/how_many_users_are_on_your_site_right_now/
I have a form (that consists of 25+ fields) and the values for these fields, range from a tiny value to a concatenated string. This is like a search tool. Now when the user fills the form and submits the info, he gets to see all the relevant data that matches the criteria in the form. I am displaying 15 records at a time to the user. I have implemented pagination to enable user to see the other records as well.
THE MAIN PROBLEM:
The part, till the users submits the info and gets back the 1st set of data is good. The problem arises when user tries to go to 2nd page (or any page of his choice) via pagination. The user is able to navigate to the other pages but the query that is needed to execute properly for pulling out the results from the DB is not triggered. Notice that initally it was a POST operation that was performed in the form and the pagination performs a GET operation. So I am losing the values of the form that the user has input and I want to retain these values and query the DB with these values.
I am trying to avoid sending the form field values via GET because I fear that the data may exceed the maximum permissible value in the URL (& as it is less secure than a POST operation). There are other operations that can be performed on the results page that can lead to loss of the form values if I try to use a POST operation (like update query). Sessions would not really work as the user can choose to run the same form in different tabs with different inputs to compare the results and this can lead to the data of the older query replaced by the data from the newer query. I haven't thought of cookies as the user may have chosen to block it. Pretty much all the options seem to be exhausted.
So what can I do to retain the form values, run the proper query and get back the relevant values irrespective of the number of times the same form may be processed by the same user in different browser tabs/windows, without using sessions(given the restrictions on passing data via GET and possibly losing them in POST operations) and be able to perform other activities on the page as well?
Thank you in advance.
First off, GET is no "less" secure than POST. Both are not to be trusted at all (It's more inviting to modify a url string, but by no means harder)...
You have a few options:
One, would be to keep a global "store" of search results. You can use a db table with id, data where data is a serialized array of the variables. Then when someone submits a search, check to see if the data is in the table. If so, use that id. If not, then serialize the data and insert it (and get the id). Then redirect to a page like: results.php?id=4. That way, they can share the link yet it stays reasonably secure from tampering (they can't alter the search params). The downside here, is that the table could grow HUGE.
Another would be to base64 encode the data and pass that through as a get parameter (base64_encode(serialize($data));). I would try to stay away from this if you're concerned with tampering or url length.
Another solution would be to intercept the next link click in JS, and use it to issue a POST back to your server from hidden variables.
EDIT: Removed the session solution. Realized that it wouldn't work for your problem.
I am trying to avoid sending the form
field values via GET because I fear
that the data may exceed the maximum
permissible value in the URL
Then you're going to have to do your pagination with a form as well, so they can also POST. Either that, or you'll have to store query terms on the server side somewhere, possibly in a session -- but don't forget that a user might have multiple tabs open at once, so you need to be able to store more than query per user.
But there's no reason you can't store multiple queries in a single session, e.g. $_SESSION['queries'][1234]. Then your pagination links would look like ?query=1234&page=3
Consider, however, that it may be useful for users to be able to share URLs of their search results. e.g., if google used POST exclusively, I couldn't send you a link to http://google.com/search?q=somequery
(& as it is less secure than a POST operation)
Not really.
A possible approach would be to store the whole query you used for searching in your database in a table say called search_queries. That table should contain essentially two columns, a hash and the query used for that search item.
When a user submits a search form, his query is evaluated and inserted in that table and he is redirected to a page with his search_hash. Every time he navigates to a different page his hash is pulled from the database and results re-evaluated accordingly -- with the proper LIMIT of course.
Make sure you cron that table (for that you might need a timestamp for each search item)
Another viable implementation of this approach would be to store the query in a SESSION variable and use it for your querying purpose. For pagination, you would /search?page=1 and your _SESSION['query'] would be for example "SELECT * FROM Topics WHERE title LIKE '%test%'". You would essentially add "LIMIT "+($page*$perpage)+", $perpage"
However the latter approach would not be able to detect multi-windows this user has with the site. You could use an array in your _SESSION['queries'] and have the user submit a /search?id=0&page=1 where id would represent which query of the array you're querying in that window.